Developmental Biology 334 (2009) 31–45

Contents lists available at ScienceDirect

Developmental Biology

j ourna l homepage: www.e lsev ie r.com/deve lopmenta lb io logy

Spatial and temporal regulation of Wnt/β-catenin signaling is essential fordevelopment of the retinal pigment epithelium

Naoko Fujimura a, Makoto M. Taketo b, Mikiro Mori c, Vladimir Korinek a, Zbynek Kozmik a,⁎a Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, Videnska 1083, 142 20 Prague 4, Czech Republicb Department of Pharmacology, Graduate School of Medicine, Kyoto University, Kyoto, 606-8501, Japanc Department of Ophthalmology, Jichi Medical School, Minami-kawachi, Tochigi, 329-0498, Japan

⁎ Corresponding author. Fax: +420 241063125.E-mail address: kozmik@img.cas.cz (Z. Kozmik).

0012-1606/$ – see front matter © 2009 Elsevier Inc. Aldoi:10.1016/j.ydbio.2009.07.002

a b s t r a c t

a r t i c l e i n f o

Article history:Received for publication 21 January 2009Revised 30 June 2009Accepted 1 July 2009Available online 9 July 2009

Keywords:Wnt/β-catenin signalingRetinal pigment epitheliumNeural retinaMitfOtx2

Wnt/β-catenin signaling is highly active in the dorsal retinal pigment epithelium (RPE) during eyedevelopment. To study the role of Wnt/β-catenin signaling in the RPE development we used a conditionalCre/loxP system in mice to inactivate or ectopically activate Wnt/β-catenin signaling in the RPE. Inactivationof Wnt/β-catenin signaling results in transdifferentiation of RPE to neural retina (NR) as documented bydownregulation of RPE-specific markers Mitf and Otx2 and ectopic expression of NR-specific markers Chx10and Rx, respectively. In contrast, ectopic activation of Wnt/β-catenin signaling results in the disruption of theRPE patterning, indicating that precise spatial and temporal regulation of Wnt/β-catenin signaling isrequired for normal RPE development. Using chromatin immunoprecipitation (ChIP) and reporter geneassays we provide evidence that Otx2 and RPE-specific isoform of Mitf, Mitf-H, are direct transcriptionaltargets of Wnt/β-catenin signaling. Combined, our data suggest that Wnt/β-catenin signaling plays anessential role in development of RPE by maintaining or inducing expression of Mitf and Otx2.

© 2009 Elsevier Inc. All rights reserved.

Introduction

The first indication of the vertebrate eye development isevagination of the diencephalon towards the surface ectoderm toform the optic vesicle (Chow and Lang, 2001). Lens-competent headectoderm responds to signals from the optic vesicle, which inducescolumnar thickening of the surface epithelium to form the lensplacode (Grainger et al., 1997). As the optic vesicle comes into contactwith the surface ectoderm, it becomes partitioned into three terri-tories: a distal territory, a proximal territory and a dorsal territory,which give rise to the neural retina (NR), the optic stalk and theretinal pigment epithelium (RPE), respectively. Coordinated invagi-nation of the optic vesicle and the lens placode leads to formation ofthe double-layered optic cup and the lens vesicle. The inner layer andthe outer layer of the optic cup give rise to the NR and RPE, res-pectively. The process of the invagination generates the optic fissurethat runs from the ventral-most region of the NR and along theventral aspect of the optic stalk. The optic fissure gradually becomesclosed and the NR is completely surrounded by the RPE. Thetransition part between the NR and the RPE called the ciliary margingives rise to the ciliary body and the iris (Bharti et al., 2006; Chowand Lang, 2001; Martinez-Morales et al., 2004).

Although little is known about the RPE development, severaltranscription factors have been shown to be involved in the process.

l rights reserved.

Mitf, Otx1, and Otx2 are essential for the RPE development, whileChx10 prevents RPE development in the presumptive NR (Horsford etal., 2005; Martinez-Morales et al., 2004; Rowan et al., 2004). Mitfencodes a member of the basic helix–loop–helix leucine zipper familyof transcription factors (Hodgkinson et al., 1993) and consists ofnine isoforms with distinct amino-termini (Hallsson et al., 2007;Steingrimsson et al., 2004). Each isoform shows a unique expressionpattern (Goding, 2000; Steingrimsson et al., 2004). For example,Mitf-A, -J, -H and -D are all expressed in the RPE, whereas expression ofMitf-M is restricted to the neural crest-derived melanocytes (Amaeet al., 1998; Bharti et al., 2008; Hershey and Fisher, 2005; Takeda et al.,2002). Mitf regulates pigment cell-specific transcription of genes en-codingmelanogenic enzymes such as tyrosinase (Tyr), and tyrosinase-related protein 1 and 2 (Aksan and Goding, 1998; Hemesath et al.,1994; Yasumoto et al., 1994, 1997). During the vertebrate eyedevelopment Mitf is expressed in the entire optic vesicle, whereaslater the expression is restricted to the RPE, the ciliary body and theiris (Baumer et al., 2003; Horsford et al., 2005; Nguyen and Arnheiter,2000). The RPE of Mitf null mutants loses the expression of RPE-specific genes and transdifferentiates into the NR (Nguyen andArnheiter, 2000). Otx1 and Otx2 encode members of the bicoid sub-family of homeodomain-containing transcription factors (Simeone etal., 1992). Similarly as Mitf, Otx1 and Otx2 are expressed in the entireoptic vesicle and later expression is restricted to the presumptive RPE(Baumer et al., 2003; Martinez-Morales et al., 2001). Otx2 cooperateswith Mitf to regulate expression of melanogenic enzymes (Martinez-Morales et al., 2003). Otx1 and 2 double-deficient mice show severe

32 N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

ocular malformation in the lens, the NR, the optic stalk and the RPE.Notably, the presumptive RPE loses expression ofMitf and gives rise tothe NR-like tissue (Martinez-Morales et al., 2001). The Chx10 geneencodes a member of paired-type homeodomain-containing tran-scription factor (Burmeister et al., 1996). Chx10 is expressed in thedistal optic vesicle and at later stages restricted to the NR progenitorcells (Baumer et al., 2003; Burmeister et al., 1996; Chen and Cepko,2000; Rowan et al., 2004). Chx10 represses expression of photo-receptor genes such as rod arrestin (Dorval et al., 2006). Chx10 nullmutant mice show expansion of the peripheral RPE into NR andectopic expression of Mitf in the entire NR (Horsford et al., 2005).Furthermore, misexpression of Chx10 in the developing RPE in chickresults in significant downregulation of Mitf and tyrosinase-relatedprotein 2, although transdifferentiation of the RPE does not occur(Rowan et al., 2004).

It has been proposed that in addition to these transcription factors,secretedmolecules from the extraocular mesenchyme are required forRPE development to inhibit the NR development in the presumptiveRPE (Fuhrmann et al., 2000). In the absence of the extraocularmesenchyme, explanted chick optic vesicles show downregulation ofRPE-specific genes and ectopic expression of NR-specific genes.Activin A, a member of the TGFβ superfamily, can substitute for theextraocular mesenchyme (Fuhrmann et al., 2000). Bone morphoge-netic proteins (BMP), other members of the TGFβ superfamily, areessential for RPE development in chick embryos. The presumptive NRdevelops into the RPE by overexpression of BMPs, while inhibition ofBMP results in abrogation of RPE development and in the induction ofexpression of the NR-specific genes (Muller et al., 2007).

In multi-cellular organisms the Wnt signaling pathway repre-sents one of the key mechanisms controlling cell-fate decisionsduring embryonic development and also in adult tissues (Klaus andBirchmeier, 2008). The signaling is initiated by the interaction ofextracellular Wnt ligands with the transmembrane Frizzled/LRPreceptor complex. The activation of the receptor results in the stabi-lization of β-catenin, a key mediator of canonical Wnt signaling. Theprotein accumulates both in the cytoplasm and the nucleus, with thenuclear form able to act as a coactivator of Tcf/Lef transcription factors(Logan and Nusse, 2004). During eye development the Wnt pathwayhas been implicated in the formationof RPE. Several components of thecanonical, i.e. β-catenin-dependent, Wnt signaling pathway, includingWnt2b, are expressed in the presumptive avian or mammalian RPE(Fuhrmann et al., 2000; Cho and Cepko, 2006; Jasoni et al., 1999; Jin etal., 2002; Liu et al., 2003; Zakin et al., 1998). In the mouse, Wnt/β-catenin signaling is highly active in the developing RPE at the stage ofthe optic cup formation and its activity is subsequently restricted to theciliary margin (Kreslova et al., 2007; Liu et al., 2006, 2007; Maretto etal., 2003; Miller et al., 2006; Zhou et al., 2008). Moreover, recentstudies have shown that Wnt/β-catenin-mediated signals are essen-tial for the ciliary margin development (Cho and Cepko, 2006; Kubo etal., 2003; Liu et al., 2007). Aberrant activation of the Wnt pathway inthe peripheral NR leads to the expansion of the ciliary margin at theexpense of the NR; on the contrary, conditional inactivation of thesignaling attenuates the ciliary margin development (Liu et al., 2007).Interestingly, β-catenin is a polypeptide with dual roles and besidesWnt signaling also participates in cell adhesion (Grigoryan et al.,2008).β-catenin directly binds to the cytoplasmic tail of cadherins andassociates withα-catenin, which links cadherin/catenin complexes tothe actin cytoskeleton (Perez-Moreno et al., 2003). Although condi-tional knockout of β-catenin results in a failure of cell adhesion andmorphogenesis in several tissues such as the presumptive lens and theforebrain (Junghans et al., 2005; Smith et al., 2005), the embryoniclethality upon complete inactivation of the β-catenin gene is morelikely related to the deficiency in the β-catenin signaling function(Huelsken et al., 2000). β-catenin-null mutant embryos preserveintact adherens junctions sinceγ-catenin, aβ-cateninparalog, possiblysubstitutes β-catenin in cell adhesion complexes in early stages of

embryonic development (Huelsken et al., 2000). A similar phenom-enon has been observed in the epidermis and cardiomyocytes(Posthaus et al., 2002; Zhou et al., 2007).

In this study we have investigated the role of the canonical Wntsignaling pathway in the RPE development by inactivating or activ-ating its key component, β-catenin. Deletion of the β-catenin generesults in transdifferentiation of the RPE to NR. On the other hand, theectopic activation of Wnt/β-catenin signaling inhibits formation ofthe RPE. We suggest that spatial and temporal regulation of Wnt/β-catenin signaling is essential for the RPE development.

Materials and methods

Mice

The following genetically modified mice were used in this study:Trp1-Cre (Mori et al., 2002), β-catenin conditional loss-of-functionCatnblox(ex2–6) (Brault et al., 2001), β-catenin gain-of-functionCatnblox(ex3) (Harada et al., 1999), Wnt/β-catenin reporter BAT-gal(Maretto et al., 2003) and Cre recombinase reporter ROSA26R(Soriano, 1999) (Jackson Laboratory, stock #0033069). The age ofmouse embryos was determined by timed pregnancy. The noon of theday onwhich the vaginal plugwas observedwas counted as embryonicday 0.5 (E0.5).

Immunohistochemistry

Mouse embryos were fixed in 4% paraformaldehyde, cryoprotectedin 30% sucrose and embedded and frozen in OCT (Tissue Tek; SakuraFinetek). Frontal frozen sectionswere prepared at 10 μmthickness. Thecryosections were re-fixed by 4% paraformaldehyde for 10 min,washed by PBS and permeabilized with 0.1% Tween 20/PBS (PBT) for15min. The sectionswere blocked by 10% BSA/PBTand incubatedwithprimary antibodies in 1% BSA/PBT overnight. The following day thesections were washed with PBS, incubated with secondary antibodiesin 1% BSA/PBT for 30 min, washed with PBS, stained with DAPI for5 min and mounted in glycerol. Immunohistochemistry was repeatedat least three times for each antibody. The following antibodies wereused in this study: anti-β-catenin (1:1000, Sigma, C2206), anti-Mitf(1:250, a gift from Dr. Arnheiter), anti-Otx2 (1:1000, a gift from Dr.Vaccarino), anti-Chx10 (1:200, Chemicon), anti-Pax6 (1:500, Cov-ance), anti-Pax2 (1:300, Zymed), anti-Nr2f1 (1:250, Perseus Proteo-mics), anti-Nr2f2 (1:250, Perseus Proteomics), anti-Sox2 (1:100, SantaCruz Biotechnology), anti-N-cadherin (1:500, BD Transduction Labo-ratories), anti-α-catenin (1/1000, Sigma), anti-γ-catenin (1:100, BDTransduction Laboratories), anti-P-cadherin (1/50, R&D Systems),anti-Lef1 (1/1000, Cell Signaling), anti-rabbit Alexa 594, anti-rabbitAlexa 488, anti-sheep Alexa 594, anti-mouse Alexa 594 and anti-goatAlexa 594 (all 1:500, Molecular Probes). Phalloidin-Alexa 488 con-jugate (1:150, Molecular Probes) was used for visualization of actinfilaments. The sections were analyzed by either Leica TCS Sp5 confocalmicroscope or Nikon DIAPHOT 300 inverted fluorescence microscope.

The β-galactosidase assay

The β-galactosidase assay was carried out as described previouslyand repeated at least three times for each genotype (Kreslova et al.,2007). Briefly, mouse embryos were fixed in 0.2% glutaraldehyde in0.1 M phosphate buffer (pH 7.3), 2 mM MgCl2 and 5 mM EGTA on ice,washed by rinse buffer (0.1 M phosphate buffer pH 7.3, 2 mM MgCl2,20mMTris–HCl pH 7.3, 0.01% sodiumdeoxycholate and 0.02% NonidetP20) and incubated in staining solution (rinse buffer supplementedwith 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide and1 mg/ml X-gal). The stained embryos were re-fixed in 4% paraformal-dehyde, cryoprotected in 30% sucrose and embedded and frozen inOCT. Frontal frozen sections were prepared at 12 μm thickness.

33N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

Whole-mount in situ hybridization

Whole-mount in situ hybridization was performed according tostandard protocols and repeated at least three times for each probe.Plasmid carryingmouse Rx (Open Biosystems) or Vax2 (a gift from Dr.Lemke) cDNA was linearized with an appropriate restriction enzyme.The antisense riboprobe was synthesized using the DIG RNA labelingkit (Roche). Mouse embryos were fixed in 4% paraformaldehyde,bleached by methanol/30% H2O2 (4:1) for 20 min and treated with20 μg/ml of Proteinase K/PBT for 15 min. The reaction was stoppedwith 2 mg/ml glycine/PBT. The embryos were washed with PBT, re-fixed with fix solution (0.2% glutaraldehyde/4% paraformaldehyde inPBS) for 20 min and washed with PBT. The embryos were incubatedwith pre-warmed prehybridization solution (50% formamide, 5× SSCpH4.5, 5mMEDTA, 0.1% Tween 20 and 50 μl/ml Heparin) for 60min at70 °C, then incubated with hybridization solution (prehybridizationsolution supplemented with 1 μg of DIG-labeled riboprobe, 25 μg/mltRNA and 25 μg/ml herring sperm DNA) overnight at 70 °C. Thefollowing day the embryos were washed with solution I (50%formamide, 4× SSC pH 4.5 and 1% SDS) at 70 °C, solution I/II (solutionII: 0.5 M NaCl, 10 mM Tris–HCl pH 7.5 and 0.1% Tween20) at 70 °C, andsolution II at RT. After the washing steps, embryos were treated with100 μg/ml RNase I at 37 °C, washed with pre-warmed solution III (50%formamide, 2× SSC pH4.5) at 65 °C, thenwith TBST (5mM Levamisole,130 mM NaCl, 2.7 mM KCl, 25 mM Tris–HCl and 0.1% Tween20) at RT.The embryos were blocked with 10% sheep serum/TBST for 2 h andincubated with anti-DIG alkaline phosphatase (1:1000, Roche) in 10%sheep serum/TBST overnight at 4 °C. Then, they were washed withTBST at RT, equilibrated with NTMT solution (100 mM NaCl, 100 mMTris–HCl pH9.5, 50 mM MgCl2 and 0.1% Tween20) and stained withNBT/BCIP (Roche) in NTMT. The stained embryos were re-fixed in4% paraformaldehyde, cryoprotected in 30% sucrose and embeddedand frozen in OCT. Frontal frozen sections were prepared at 12 μmthickness.

Chromatin immunoprecipitation (ChIP)

The chromatin immunoprecipitation assay was performed as des-cribed previously (Fujimura et al., 2007). Briefly, embryos at E10.5were homogenized and cross-linked in 1% formaldehyde in PBS for10 min at room temperature. The cross-linking reaction was stoppedby adding glycine. Cross-linked cells werewashedwith PBS containingproteinase inhibitors and sonicated in 2× SDS buffer (1% SDS, 10 mMEDTA, 50 mM Tris–HCl, pH 8.0). Approximately 30 μg of sonicatedchromatinwere diluted ten times with dilution buffer (0.01% SDS,1.1%Triton X-100, 1.2 mM EDTA, 16.7 mM Tris–HCl, pH 8.0, 167 mM NaCl)containing protein inhibitors, then precleared with 50 μl of protein Aagarose slurry (Upstate Biotech) for 3 h. The precleared chromatinwasincubated either with 20 μg of anti-β-catenin rabbit serum, anti-Tcf/Lef rabbit serum (Valenta et al., 2006) or with a control IgG overnightat 4 °C (Valenta et al., 2006). Protein A agarose slurry (20 μl) wasadded and incubated for 3 h at 4 °C. The samples were washed threetimes in low-salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA,20 mM Tris–HCl, pH 8.0, 150 mMNaCl), three times in high-salt buffer(0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris–HCl, pH 8.0,500 mM NaCl), four times in LiCl buffer (0.25 M LiCl, 1% Nonidet P-40,1% deoxycholate, 1 mM EDTA, 10mM Tris–HCl, pH 8.0) and twice in TEbuffer (10 mM Tris–HCl, 1 mM EDTA, pH 8.0). After washing, theimmunocomplex was eluted twice using 100 μl of elution buffer(1% SDS, 1 mM EDTA, 10 mM Tris–HCl, pH 8.0) for 20 min at roomtemperature. Immunoprecipitated DNA was de-cross-linked over-night at 65 °C in the presence of proteinase K (Roche) and 250 mMNaCl. The samples were purified using MinElute reaction cleanup kit(Qiagen) with a final elution volume of 50 μl. One μl of the purifiedsample was used for PCR. The ChIP assay was repeated three times.The primers used were as follows (written in the 5′ to 3′ direction):

Mitf-H, GGCTAAACTCCCTGGCTGAC and CCTTGGCGAAGAGTGAAAGT;Mitf-D, TTGCGCTATCCCCTTTAGAA and GCTTCATGCTTCAACCACAA;Nr2f2-pro, AGGCATGAAATAGGGGAACC and TCCTCTCACCCTGGAT-CAAA; Nr2f2-UTR, AGTGTGCAGGCTTTTCCAAC and CTTTAGTT-TGGCGGGTGAAA; Nr2f2-ECR1, TTGGATTTTAGGCAGCAAGG andAGAGAAGGGGCTTCCATTGT; Nr2f2-ECR2, GCTAAGTTGCAGCAGTCGTGand GGGGAGAAACGGAGAGAAAG; Nr2f2-ECR3, GAGGGGAAAGAAA-CAGCTCA and TACAGTGTGCATGGGGATTG; Otx2-FM2, CTCTCACTAC-CCCCACGAGA and TCACCGTTCGGAGATAATCC.

Cell culture, transient transfection, and luciferase reporter assay

ARPE19 cells were cultured in Dulbecco's modified Eagle's medium(DMEM)/F-12 Ham mixture [supplemented with 10% fetal bovineserum (PAA Laboratories), 2mM L-glutamine,100 U/ml penicillin, and0.1 mg/ml streptomycin (Sigma)]. Transient transfection of ARPE19cells was performed in triplicates using FuGENE 6 (Roche AppliedScience) according to themanufacturer's protocol. Cells were plated in24-well plates 24 h prior to transfection. Typically, the total amount ofDNA transfected per well was 300 ng and was adjusted with pUC 18when necessary. A Renilla luciferase expression plasmid was cotrans-fected to normalize the transfection efficiency. Two days aftertransfection, the cells were lysed in 100 μl of passive lysis buffer andluciferase reporter assays were performed using Dual-LuciferaseReporter assay system (Promega).

The mouse Mitf-H promoter, Mitf-D promoter, Otx2 FM2 enhancer,Nr2f2 promoter Nr2f2 5′ UTR, and Nr2f2 ECR1 were amplified by PCRusing C57BL/6J mouse genomic DNA as a template. Promoter andenhancer fragments were cloned into pGL3 basic or pGL4.26 vector,respectively. Each luciferase reporter construct (200 ng) wascotransfected with the stabilized form of β-catenin (β-cat) and Lef1(50 ng each) or an empty vector (100 ng) into ARPE19 cells.

Results

Conditional inactivation of Wnt/β-catenin signaling in the presumptiveretinal pigment epithelium

Mouse strains with the integrated Wnt/β-catenin-responsivereporters represent a valuable tool to monitor the activity of canonicalWnt signaling during embryogenesis and in adult tissues (DasGuptaand Fuchs, 1999; Maretto et al., 2003; Mohamed et al., 2004). Usingsuch “reporter mice” we and others have previously shown that Wntsignaling is active during eye development (Kreslova et al., 2007; Liuet al., 2006; Maretto et al., 2003; Smith et al., 2005). In this study, weassessed the role of the canonical Wnt pathway in development of theRPE using transgenic mice containing the lacZ gene regulated by asynthetic enhancer element composed of seven optimal Tcf/Lefbinding DNA motifs (BAT-gal) (Maretto et al., 2003). As indicated bythe lacZ expression, Wnt/β-catenin signaling is highly active in thepresumptive RPE at E10.5, and subsequently, the lacZ activity isrestricted to the distal RPE (Figs. 1A, C) (Kreslova et al., 2007; Liu et al.,2006). Importantly, Axin2 and Lef1, direct target genes of Wnt/β-catenin signaling, and Wnt2b are expressed in the distal RPE(Supplementary Figs. 1A, B) (Burns et al., 2008; Filali et al., 2002;Hovanes et al., 2001; Jho et al., 2002; Liu et al., 2006; Yan et al., 2001;Zakin et al., 1998). Based on these data, we hypothesized that Wnt/β-catenin signaling plays a role in the RPE development. To test thishypothesis, we manipulated the Wnt pathway in the RPE byconditional inactivation or activation of β-catenin using the Cre/loxPsystem inmice.We utilized the Trp1-Cre transgenic mice as a “deleter”(Mori et al., 2002). In these mice Cre recombinase is active in the RPEfrom E10.0 and its expression is also detected in the optic stalk andthe ciliary margin at a later stage (Mori et al., 2002) (SupplementaryFigs. 2A, B). To achieve conditional inactivation of Wnt/β-cateninsignaling, Trp1-Cre was combined with a conditional allele of β-

Fig. 1. Conditional inactivation of Wnt/β-catenin signaling in the RPE. The activity of Wnt/β-catenin signaling was assessed using Wnt/β-catenin signaling reporter line BAT-gal.Whole-mount or frontal cryosections of wild-type (A, C) and loss-of-function mutant (B, D) were stained with X-gal at the indicated embryonic stage. Confocal images of frontalcryosections of wild-type (E, E′) and loss-of-function mutant (F, F′) showing expression of β-catenin (red) and DAPI nuclear labeling (blue) at E10.5. Small areas of RPE in panels(E, E′) and (F, F′) are shown magnified below the corresponding panel. The RPE is indicated by a dashed line. Abbreviations used in this and subsequent figures are as follows: rpe,retinal pigment epithelium (RPE); le, lens; nr, neural retina (NR).

34 N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

catenin Catnblox(ex2–6) in which exons 2–6 are flanked by loxP sites(Brault et al., 2001). Since β-catenin acts as a coactivator of Tcf/Leftranscription factors (Logan and Nusse, 2004), Cre-mediated condi-tional deletion of β-catenin results in inactivation of Wnt/β-cateninsignaling. The genotype of each embryo used in this study is shown inthe corresponding figure panel; however, for the sake of simplicity, we

refer only to Catnblox(ex2–6)/lox(ex2–6) as wild-type and Trp1-Cre;Catnblox(ex2–6)/lox(ex2–6) as loss-of-function mutants in the text.

We assessed the β-catenin levels in the RPE of loss-of-functionmutant using immunohistochemical staining and confirmed that theamounts of β-catenin protein in the RPE was reduced at E10.5(Figs.1E, E′, F, F′). We further investigated the status ofWnt/β-catenin

35N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

signaling in the presumptive RPE of loss-of-function mutants hemi-zygous for the BAT-gal reporter. At E10.5 and E11.0, the reporteractivity was detected in the presumptive RPE of wild-type (Figs. 1A,C); on the contrary, in the loss-of-function mutants the lacZ reporteractivity was significantly decreased (Figs. 1B, D). In addition, Lef1 wasdownregulated in the presumptive RPE upon loss of the β-cateningene (Supplementary Fig. 1C). These results imply that Wnt/β-catenin signaling is active in the dorsal RPE and becomes inactivatedin the loss-of-function mutants.

Wnt/β-catenin signaling is required for the RPE development

The phenotypic consequences of the inactivation of Wnt/β-catenin signaling in the RPE were further investigated using histo-chemistry. Abnormal development of the RPE was observed at E10.5.The prospective RPE of the wild-type mice was thin and the cellsformed a single layer (Fig. 2A); on the contrary, the analogous tissue ofloss-of-function mutants was thicker (Fig. 2B). Moreover, the RPE washypopigmented and the ventral NR was shortened (Figs. 2E–J).Furthermore, the thickening of the ventral RPE was also observedduring later stages of the embryonic development (Figs. 2E–H).

It is well established that β-catenin functions in both cell adhesionandWnt/β-catenin signaling. Therefore, we tested the possibility thatthe malformations of the RPE are connected with the reduced or

Fig. 2. Phenotypic consequences in the loss-of-function β-catenin mutant. Frontal cryosectioby hematoxylin and eosin at the indicated embryonic stage. Red lines indicate the size of thearrowheads (in E, G). The embryo heads show the loss of pigmentation in the loss-of-funct

absent cell-to-cell contacts. We performed immunohistochemicalanalysis of the adhesion-related proteins N-cadherin, α-catenin andγ-catenin. In addition, phalloidin staining was used to detect possibleabnormalities of actin filaments. At E10.5, N-cadherin is expressed inall retinal cells (Xu et al., 2002). γ-catenin is known to substitute forβ-catenin in adherens junctions under conditions when β-catenin isabsent or modified (Huelsken et al., 2000; Posthaus et al., 2002; Zhouet al., 2007). Confocal microscopy images showed identical stainingpattern of F-actin, N-cadherin, γ-catenin and α-catenin in the wild-type and β-catenin-deficient RPE (Figs. 3A–L, Supplementary Figs. 3A,B), suggesting that cellular adhesion is not affected by the loss ofβ-catenin. Altogether, our data suggest that cell adhesion is not grosslyaffected in the absence of β-catenin at E10.5. Thus, the defect in theRPE development is likely to be caused by the absence of the β-cateninsignaling function.

To characterize the phenotype caused by inactivation of Wnt/β-catenin signaling we first examined the expression of Mitf and Otx2,i.e. the genes that are essential for the RPE development (Bharti et al.,2006; Goding, 2000; Hodgkinson et al., 1993; Martinez-Morales et al.,2004; Martinez-Morales et al., 2001). Mitf and Otx2 are initiallyexpressed throughout the optic vesicle; subsequently their expressionis restricted to the dorsal optic vesicle, which gives rise to the RPE(Martinez-Morales et al., 2001; Nguyen and Arnheiter, 2000). At E10.5,both Mitf and Otx2 were detected in the wild-type RPE (Figs. 4A, C).

ns of wild-type (A, C, E, and G) and loss-of-function mutant (B, D, F, and H) were stainedventral and dorsal NR, respectively. The presumptive RPE is indicated by a dashed line orion mutant (I, J).

Fig. 3. Conditional inactivation of β-catenin in the RPE does not affect cell adhesion. Frontal cryosections of wild-type (A, C, and E) and loss-of-function mutant (B, D, and F) wereimmunostained with β-catenin and phalloidin (A, B), β-catenin and N-cadherin (C, D), or β-catenin and γ-catenin (E, F) antibodies. Magnified views of RPE in panels (A–F) are shownnext to the corresponding panel (G–L). The images were obtained by confocal microscope analysis. The RPE is indicated by a dashed line.

36 N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

Fig. 4. Conditional inactivation of Wnt/β-catenin signaling results in transdifferentiation of RPE to NR. Frontal cryosections of wild-type (A, C, E and G) and loss-of-function mutant(B, D, F and H) at E10.5 were immunostained with Mitf (A, B), Otx2 (C, D) and Chx10 (E, F) antibodies or hybridized with an antisense probe against Rx (G, H). The RPE is indicated bya dashed line or an arrowhead. Note that the RPE-specific markers Mitf and Otx2 are significantly downregulated and the NR-specific markers Chx10 and Rx are ectopically expressedin the thickened presumptive RPE of the loss-of-function mutant.

37N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

Interestingly, expression of Mitf and Otx2 was significantly down-regulated in the thickened RPE of the loss-of-function mutant (Figs.4B, D). Likewise, the expression of RPE-specific P-cadherin (Xu et al.,2002) was lost in the thickened RPE of the loss-of-function mutant(Supplementary Figs. 3C, D). The absence of Mitf results intransdifferentiation of the RPE into the NR (Nguyen and Arnheiter,2000). Thus, we investigated whether the RPE in the loss-of-functionanimals expresses Chx10 and Rx that are essential for the NRdevelopment (Baumer et al., 2003; Burmeister et al., 1996; Matherset al., 1997). As shown in Figs. 4E–H, in contrast to the wild-type micethe expression of Chx10 and Rx in the loss-of function mutants wasnot limited to the NR but also extended to the presumptive RPE. Theseresults indicate that Wnt/β-catenin signaling controls the cell-fatedecision in the developing RPE and in its absence the tissuetransdifferentiates into the NR.

Ectopic activation of Wnt/ β-catenin signaling in the RPE

As evidenced by the activity of the BAT-gal reporter and theexpression pattern of the Axin2 and Lef1 genes the canonical Wntpathway is active in the dorsal part of the presumptive RPE at E10.5(Fig. 1, Supplementary Fig. 1). Subsequently, its activity is restricted tothe peripheral RPE (Kreslova et al., 2007). Based on these data, wehypothesized that the restricted pattern of Wnt/β-catenin signalingmight be required and that ectopic activation of the signaling in theentire RPE could impair the RPE development. To achieve the ectopicactivation of Wnt/β-catenin signaling in the entire RPE, Trp1-Cremicewere crossed with Catnblox(ex3) mice in which exon 3 of the β-cateningene is flanked by loxP sites (Harada et al., 1999). Exon 3 encodes Ser/Thr residues that are phosphorylated in the absence of Wnt signalingand the entire protein is subsequently degraded by proteasome. Cre-mediated recombination results in the production of a stabilized form

of β-catenin protein. This protein variant accumulates in the cells andconstitutively activates Wnt signaling. The genotype of each embryoused in the analysis is shown in the corresponding figure panel;however, for the sake of simplicity, we refer only to Catnblox(ex3)/+ aswild-type and Trp1-Cre; Catnblox(ex3)/+ as gain-of-function mutant inthe text.

In order to assess the level of β-catenin in the RPE of gain-of-function mutants, we performed immunohistochemistry using β-catenin antibody recognizing the C-terminus of the protein. Incontrast to the wild-type mice high levels of β-catenin were detectedin the entire RPE at E11.5 and in the proximal part of the disorganizedretina at E13.5 (Figs. 5E–H). Of note, nuclear β-catenin was detectedat both stages (Figs. 5K, L). The activity of the BAT-gal reporter wasclearly detected in the cells with increased β-catenin levels. More-over, the area in the RPE that showed the reporter activity expandedventrally in the gain-of-function mutant embryos at E10.5 and E11.5(Figs. 5A–D).

Ectopic activation of Wnt/β-catenin signaling disrupts development ofthe RPE

At E10.5 the entire RPE of the gain-of-function mutant becamethicker compared to that of wild-type (Figs. 6A, B). Interestingly, theproximal part of the RPE was significantly thickened while the distalpart of the RPE remained thin by E11.5 (Figs. 6C, D). At later stages, theentire retinawas disorganized (Figs. 6E–H) and the embryos showed acomplete loss of pigmentation (Figs. 6I, J).

To assess the differentiation of RPE and NR, we examinedexpression of RPE-specific markers Otx2 and Mitf, and NR-specificmarkers Chx10 and Rx (Bharti et al., 2006; Chow and Lang, 2001;Martinez-Morales et al., 2004). At E10.5 the expression of Otx2, Mitf,Chx10 and Rxwas not changed (data not shown) despite the morpho-

Fig. 5. Ectopic activation of Wnt/β-catenin signaling in the RPE. The activity of Wnt/β-catenin signaling was assessed using Wnt/β-catenin signaling reporter line BAT-gal. Whole-mount or frontal cryosections of wild-type (A, C) and gain-of-function mutant (B, D) were stained with X-gal to assess the activity of Wnt/β-catenin signaling at the indicatedembryonic stage. The extent of the RPE area in which Wnt/β-catenin signaling is active is marked by arrowheads. Confocal images of frontal cryosections fromwild-type (E, G) andgain-of-function mutant (F, H) showing expression of β-catenin (red) and DAPI nuclear labeling (blue) at E10.5. Small areas of RPE in panels (E, G) and (F, H) are shownmagnified inpanels (I, J) and (K, L), respectively. The RPE layer of wild-type and the presumptive RPE tissue of themutant (marked by the increased level of stabilized β-catenin) are indicated by adashed line.

38 N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

Fig. 6. Phenotypic consequences in the gain-of-function β-catenin mutant. Frontal cryosections of wild-type (A, C, E and G) and gain-of-function mutant (B, D, F and H) were stainedby hematoxylin and eosin at the indicated embryonic stage. The RPE is indicated by a dashed line or an arrowhead. The embryo heads show the loss of pigmentation in the gain-of-function mutant (I, J).

39N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

logical differences in the RPE of the gain-of-functionmice (Figs. 6A, B).By E11.5, the expression of Mitf and Otx2 was downregulated in thepresumptive RPE of gain-of-function mutant; expression of bothgenes appeared to be only detected in the distal part of the RPE andnotably absent in the proximal part, inwhichWnt/β-catenin signalingis ectopically activated (Figs. 7A–D). On the contrary, there was nosignificant difference in the expression pattern of the NR markersChx10 and Rx (Figs. 7E–H). This suggests that in contrast to thephenotypic changes in the loss-of-function mice the abnormalthickness of the proximal RPE is not caused by transdifferentiationto NR. Next we examined the expression pattern of Pax6 and Pax2,which are involved in the RPE development (Baumer et al., 2003).Pax6 is initially detected throughout the optic vesicle and itsexpression is later maintained in both NR and RPE (Ashery-Padan etal., 2000; Baumer et al., 2003; Marquardt et al., 2001). Null mutationsof Pax6 in mice result in failure of the optic vesicle formation(Grindley et al., 1995). Ectopic expression of Pax6 in the mouse opticstalk under the control of Pax2 promoter results in expansion of theRPE, suggesting that Pax6 is a positive regulator of RPE (Baumer et al.,2003). Pax2 is initially detected in the entire optic vesicle; subse-quently the expression is restricted to the ventral optic cup and theoptic stalk (Baumer et al., 2003; Nornes et al., 1990; Schwarz et al.,2000; Torres et al., 1996). A null mutation of Pax2 results in extensionof the RPE into the optic stalk, suggesting that Pax2 is essential for

formation of the RPE/optic stalk boundary by suppressing the RPEdevelopment at the boundary (Torres et al., 1996). As shown in Fig. 7,Pax6 was downregulated in the proximal RPE of gain-of-functionmutant at E11.5 (compare panels 7I and 7J). In order to obtain moreinformation about the effects of activated Wnt/β-catenin signaling onthe distal/proximal axis of RPE we performed immunohistochemistrywith horizontal sections at E11.5. There was no significant differencein the Pax2 expression pattern in wild-type and gain-of-functionmutant embryos at E11.5 and E13.5, indicating that the optic stalkdevelopment was not affected by ectopic activation of Wnt signaling(Supplementary Figs. 4I, J and Figs. 7S, T). The expression of Pax6, Mitfand Otx2 was downregulated in the entire RPE of the gain-of-functionmutants, while the expression of Chx10, a marker of the NR, virtuallydid not differ from the wild embryos (Supplementary Fig. 4). At E13.5,the expression of Mitf and Otx2 remained undetectable in the gain-of-function mutants (Figs. 7K–N), whereas a production of the neuralretinamarkers Sox2 and Chx10was clearly localized to the abnormallyfolded neural retina but not to the presumptive RPE (Figs. 7O–R).Thus, none of the relevant markers was detected in cells expressingstabilized β-catenin. Taken together, our results indicate that thespatial and temporal activation or restriction of Wnt/β-cateninsignaling is essential for proper development of the RPE. Nevertheless,the developmental fate of the cells resulting from aberrant Wntsignaling remains elusive.

Fig. 7. Ectopic activation of Wnt/β-catenin signaling results in the abrogation of the RPE development. Frontal cryosections of wild-type (A, C, E, G, I, K, M, O, Q, and S) and gain-of-function mutant (B, D, F, H, J, L, N, P, R, and T) at E11.5 or E13.5 were immunostained with Mitf (A, B, K, and L), Otx2 (C, D, M, and N), Chx10 (E, F, O, and P), Pax6 (I, J), Sox2 (Q, R), Pax2(S, T) antibodies or hybridized with an antisense probe against Rx (G and H). The RPE is indicated by a dashed line or an arrowhead. Note that Mitf and Otx2 are downregulatedspecifically in the proximal RPE of the gain-of-function mutant at E11.5 and E13.5.

40 N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

Wnt/β-catenin signaling directly regulates expression of the genes thatare essential for the RPE development

The loss-of-function mutant exhibited transdifferentiation of RPEto NR, suggesting that Wnt/β-catenin signaling could suppressneurogenesis in the RPE. In accordance with this hypothesis it hasbeen shown that ectopic activation of Wnt/β-catenin signaling in thecentral retina results in loss of NR identity (Fu et al., 2006). One of thepossible explanations is that Wnt signaling induces expression oftranscriptional repressors or corepressors to inhibit transcription ofthe NR markers in the RPE. The identity of such repressive factors isunknown. Several RPE-expressed factors such as Pax6, Pax2, Vax2,

Nr2f1 (COUP-TFI) and Nr2f2 (COUP-TFII) were previously shown toact as transcriptional repressors (Cai et al., 2003; Duncan et al., 1998;Mui et al., 2005; Naka et al., 2008). Vax2 encodes a homeobox-containing transcriptional repressor and is expressed in the ventraloptic cup (Barbieri et al., 1999; Kim and Lemke, 2006; Mui et al., 2005;Schulte et al., 1999). In Vax1 and Vax2 double-knockout mice, thedorsal RPE expands into the dorsal optic stalk and development of theventral RPE is disrupted (Mui et al., 2005). Nr2f1 and Nr2f2 belongs toorphan members of the steroid/thyroid hormone receptor super-family (Cooney et al., 1992; Pereira et al., 2000). Nr2f2 is expressed inthe RPE, the dorsal NR, and the optic stalk during eye development(McCaffery et al., 1999). Despite its expression in the developing eye

41N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

the function of Nr2f2 during eye development is not known since theinactivation of the Nr2f2 gene in the mouse is lethal around E10.0(Pereira et al., 1999). Nr2f1 and Nr2f2 have been shown to antagonizethe nuclear receptor signaling pathway (Beland and Lohnes, 2005;Butler and Parker, 1995; Cooney et al., 1992).

We examined the expression of the aforementioned transcrip-tional factors in loss-of-function mutants at E10.5 using in situhybridization and immunohistochemistry. There was no obviousdifference in the expression pattern of Pax2 and Pax6 between thewild-type and loss-of-function animals (Supplementary Figs. 5A–D).However, we observed changes in the expression of the Vax2 gene.While in the wild-type embryos Vax2 mRNA was detected in theventral parts of the NR and RPE, in the transdifferentiated RPE the areaof Vax2 expression extended proximally (Supplementary Figs. 5I, J).We hypothesized that the ectopic Vax2 expression could be due to thedisturbed dorsoventral polarity in the transdifferentiated RPE. Nr2f1was detected in the ventral NR, the ventral RPE, and the optic stalk ofthe wild-typemice (Supplementary Fig. 5E). In the transdifferentiatedRPE of the loss-of-function mutants the expression of Nr2f1 expandeddistally (Supplementary Fig. 5F), presumably for the same reason asthat of Vax2. Nr2f2 was expressed in the distal NR, the optic stalk, andthe RPE (Supplementary Fig. 5G). Interestingly, Nr2f2 was down-regulated in the transdifferentiated RPE of loss-of-function mutants(Supplementary Fig. 5H), suggesting that Nr2f2 is regulated by Wnt/β-catenin signaling and could be involved in inhibition of the neuraldifferentiation in the RPE.

As shown above, Mitf and Otx2 were downregulated in thetransdifferentiated RPE of loss-of-function mutants, indicating thatMitf and Otx2 may be directly regulated by Wnt/β-catenin signaling.Wnt/β-catenin signaling-mediated regulation of Otx2 and Mitf has

Fig. 8. Wnt/β-catenin signaling directly regulates Mitf and Otx2 genes. (A) Map of the Mitdepicted by black ovals. (B) Chromatin immunoprecipitationwas performed with wild-type especific primers detecting the indicated transcriptional regulatory elements described in panMitf-D, and the enhancer of Otx2. (C) The indicated luciferase reporter plasmids (200 ng) werempty expression plasmid (100 ng) into ARPE19 cells. Renilla luciferase expression plasmid

been investigated (Dorsky et al., 2000; Kurokawa et al., 2004; Takedaet al., 2000). The Otx2 enhancer, which contains multiple Tcf/Lefbinding sites, was found to be active in RPE as well as in forebrain andmidbrain; mutations in Tcf/Lef sites nearly abolish the enhanceractivity, indicating that the expression of Otx2 is directly regulated byWnt/β-catenin signaling (Kurokawa et al., 2004). Among Mitfisoforms, Mitf-M is known to be directly regulated by Wnt/β-cateninsignaling (Takeda et al., 2000). However, Mitf-M is melanocyte-specific and is not expressed in RPE (Bharti et al., 2008). Recent studyhas shown that two other isoforms, Mitf-H and Mitf-D, play a key rolein the RPE development (Bharti et al., 2008). Interestingly, theexpression of Mitf-D and Mitf-H is initiated in the presumptive RPE atE10.5 (Bharti et al., 2008), which coincides with initiation of Wnt/β-catenin signaling in the presumptive RPE (Kreslova et al., 2007). Toinvestigate possible direct regulation by Wnt/β-catenin signaling wesearched for Tcf/Lef binding sites within the promoters of Mitf-D,Mitf-H, and Nr2f2. We found that the promoters of Mitf-D, Mitf-H, andNr2f2 as well as 5′ UTR of Nr2f2 contain potential Tcf/Lef binding sites(Fig. 8A and Supplementary Fig. 5K). Additionally, we comparedsequences of Nr2f2 introns of frogs, chickens, humans and mice in anattempt to find other potential transcriptional regulatory elementssince these elements are often evolutionarily conserved. We foundthree evolutionarily conserved regions containing potential Tcf/Lefbinding sites that are located 1.8 kb, 5.1 kb, and 11 kb downstream ofthe Nr2f2 promoter, referred to as ECR1, ECR2, and ECR3, respectively(Supplementary Fig. 5K). In order to examine whether these bindingsites are responsible for mediatingWnt/β-catenin signaling through adirect interaction with the Tcf/Lef/β-catenin complex, chromatinimmunoprecipitation was performed with β-catenin and Tcf/Lefantibody using wild-type embryos at E10.5. The previously described

f-D, Mitf-H, and Otx2 (Kurokawa et al., 2004) locus with putative Tcf/Lef binding sitesmbryos at E10.5 using Tcf/Lef, β-catenin or control antiserum. PCR was carried out withel (A). The presence of β-catenin and Tcf/Lef was confirmed on the promoters ofMitf-H,e cotransfectedwith the stabilized form of β-catenin (β-cat) and LEF1 (50 ng each) or an(10 ng) was cotransfected to normalize for transfection efficiency.

42 N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

Otx2 enhancer (Kurokawa et al., 2004) was also analyzed by chromatinimmunoprecipitation to confirm that the enhancer is regulated byWnt/β-catenin signaling. We found that β-catenin and Tcf/Lef werepresent on the potential Wnt/β-catenin signaling-responsive ele-ments of Nr2f2, Mitf-D and Mitf-H, with the exception of ECR2 andECR3 (Fig. 8B, Supplementary Fig. 5). We further investigated whetherthese elements are Wnt/β-catenin signaling-responsive by theluciferase reporter gene assay. Lef1/β-catenin activated the Mitf-Hpromoter and Otx2 FM2 enhancer (Fig. 8C), but not the other putativeregulatory elements (data not shown). These results indicate thatWnt/β-catenin signaling directly regulates expression of Mitf-H andOtx2, and possibly, of Mitf-D and Nr2f2. Taken together, our datasuggest that Wnt/β-catenin signaling regulates the RPE developmentby inducing expression of the key differentiation genes.

Discussion

In this study we have investigated the role of Wnt/β-cateninsignaling in the RPE development using conditional loss- and gain-of-function β-catenin alleles. Intriguingly, elimination of β-catenin in theRPE resulted in transdifferentiation of the dorsal RPE to NR asevidenced by downregulation of the RPE-specific markers Mitf andOtx2 and ectopic expression of the NR-specific genes Chx10 and Rx.Given the dual role of β-catenin in transcription activation and in celladhesion we first assessed which of these molecular functions con-tribute to the observed phenotype.

The fact that the absence of canonicalWnt signaling upon loss of β-catenin was the primary mechanism of the aberrant development ofRPE was concluded from the following observations. We and othershave shown that during mouse eye development Wnt signaling isinitially detected in the distal part of the presumptive RPE. In the laterdevelopmental stages the active Wnt pathway is restricted to theperipheral area of the RPE [Fig. 1 and results in references (Kreslova etal., 2007; Liu et al., 2006)]. As all these studies used two well-established but nevertheless synthetic Wnt/β-catenin signalingreporter genes, along with the β-galactosidase assay we performedhybridization of mRNA in situ. This analysis clearly showed that theactivity of the reporters in the RPE correlates well with the expressionpattern of two endogenous target genes of canonical Wnt signaling,Axin2 and Lef1 (Supplementary Figs. 1A, B). In addition, active Wnt/β-catenin signaling is detected in the presumptive RPE not only inmice but also in chickens, fishes, and frogs (Dorsky et al., 2002; Cho

Fig. 9. A model summarizing the role of Wnt/β-catenin signaling in RPE development. (A) Tsignaling in the RPE directly maintains or induces expression of Mitf-H, Mitf-D, Otx2 andtyrosinase (Aksan and Goding, 1998; Martinez-Morales et al., 2003). Nr2f2 may positively reNR-specific genes. In NR the expression of Chx10 and Rx is maintained, and Chx10 directly inhal., 2004). Dashed lines indicate indirect regulation, whereas solid lines indicate direct regregulated by Wnt/β-catenin signaling. On the contrary, the ventral RPE is regulated by Son

and Cepko, 2006; Kreslova et al., 2007; Liu et al., 2006; Van Raay et al.,2005; Veien et al., 2008). Available data thus suggest an ancient andprominent role for Wnt/β-catenin signaling in the control of RPEdevelopment. The expression of both BAT-gal reporter and Lef1 geneswas significantly downregulated in the presumptive RPE of the loss-of-function β-catenin animals (Figs. 1B, D and Supplementary Fig. 1C),indicating thatWnt signaling is indeed disrupted in themutant RPE. Incontrast to Wnt signaling, the loss of β-catenin did not influence celladhesion in the affected tissues as evidenced by the overall cellmorphology and unchanged expression and localization of actinfilaments and membrane N-cadherin (Fig. 3). We concluded that thelack of β-catenin is compensated for by its paralog γ-catenin asobserved previously in various developmental contexts (Huelsken etal., 2000; Posthaus et al., 2002; Zhou et al., 2007). Intriguingly, in theNR β-catenin does not regulate cell differentiation but is essential forthe normal migration. In β-catenin-null neural retinas Fu et al.detected retinal lamination defects possibly related to disorganizedlocalization of N-cadherin and F-actin and decreased cell adhesion (Fuet al., 2006). Why β-catenin is differentially involved in signaling(RPE) or adhesion (NR) in the different parts of retina and why the β-catenin function in the NR cannot be compensated by γ-cateninremains elusive. The RPE-to-NR transdifferentiation is reminiscent ofthe phenotype described in Mitf-deficient and Otx1/Otx2 double-deficient mice (Martinez-Morales et al., 2001; Nguyen and Arnheiter,2000). Interestingly, Otx2 is directly regulated by Wnt/β-cateninsignaling via its remote enhancer [Fig. 8 and see also reference(Kurokawa et al., 2004)] and furthermore, the promoter of Mitf-H, theRPE-enriched isoform of the Mitf gene, responds to the Lef/β-cateninstimulation (Fig. 8). Altogether, the published data and experimentalevidence presented in this study strongly support the idea that Wnt/β-catenin signaling is required for the proper RPE specification. Wesuggest that canonical Wnt signaling induces or maintains theexpression of the RPE-specific transcription factors and simulta-neously inhibits the NR differentiation program in the developing RPE.

It is worth noting that the expression of melanogenic genes such asTyr is induced at E10.5 (Beermann et al., 1992), which coincides withinitiation of Wnt/β-catenin signaling in the presumptive RPE.Expression of melanogenic genes is directly regulated by thetranscription factors Mitf and Otx2 (Aksan and Goding, 1998;Hemesath et al., 1994; Martinez-Morales et al., 2003; Yasumoto etal., 1994, 1997) (Fig. 9A). Even though Mitf and Otx2 are expressed inthe entire optic vesicle at E9.5 (Baumer et al., 2003; Martinez-Morales

ranscriptional network downstream of Wnt/β-catenin signaling in RPE. Wnt/β-cateninNr2f2. Mitf-D, Mitf-H and Otx2 regulate expression of the melanogenic genes such asgulate expression of genes that are involved in RPE development or inhibit expression ofibits expression ofMitf-D andMitf-H (Bharti et al., 2008; Horsford et al., 2005; Rowan etulation. (B) Signaling pathways in developing RPE. The distal part of the dorsal RPE isic hedgehog (Shh) signaling (Zhang and Yang, 2001).

43N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

et al., 2001; Nguyen and Arnheiter, 2000), the optic vesicle of Mitf-and Otx1/Otx2-deficient mice is normal, and their target genes such asTyr are not expressed (Martinez-Morales et al., 2001; Nguyen andArnheiter, 2000). This indicates that their transcriptional activitymight be regulated by an RPE-specific signal. Interestingly, recentstudies have indicated that Wnt/β-catenin signaling modifies tran-scriptional properties of the Mitf protein. First of all, Mitf directlyinteracts with β-catenin and recruits β-catenin on the its targetpromoters as a transcriptional activator in melanoma cells (Schepskyet al., 2006). Another study has suggested that the transcriptionalactivity of Mitf is attenuated until Pax3-mediated repression isrelieved by Wnt/β-catenin signaling in melanocyte stem cells (Langet al., 2005). Thus, it is possible that Wnt/β-catenin signaling plays arole as a differentiation signal within the RPE by boosting thetranscriptional activity of Mitf (and possibly Otx2) in a similar manneras in melanocyte stem cells (Fig. 9A).

Wnt-mediated inhibition of neural differentiation has beenimplicated in the ciliary margin development (Cho and Cepko,2006; Liu et al., 2007). Ectopic activation of Wnt/β-catenin signalingin the peripheral NR results in expansion of the ciliary margin towardthe NR and inhibition of neurogenesis (Liu et al., 2007). On thecontrary, conditional inactivation of the signaling in the peripheral NRleads to downregulation of the ciliary margin markers Otx1 and Msx1(Liu et al., 2007). Similarly, ectopic activation of Wnt/β-cateninsignaling in the chick eye inhibits the NR differentiation and inducesthe ciliary margin markers (Cho and Cepko, 2006). It is thereforepossible that one of the roles of Wnt/β-catenin signaling in the RPE isto inhibit neurogenesis in a similar manner as in the ciliary margin.

Interestingly, in contrast to the dorsal RPE, the ventral RPEremained thin and contained a subset of pigmented cells until E13.5in the loss-of-function β-catenin mutant. This may suggest that thedorsal and the ventral RPE are regulated by two distinct mechanisms.Supporting this idea, in AP2α-deficient mice only the dorsal but notthe ventral RPE transdifferentiates into NR (West-Mays et al., 1999). Incontrast, only the ventral RPE-to-NR transdifferentiation occurs in nullmutantmice of Gas1, which positively regulates Sonic hedgehog (Shh)signaling (Allen et al., 2007; Lee et al., 2001; Martinelli and Fan, 2007).Shh signaling is important for the ventral specification of the eye(Chow and Lang, 2001; Mui et al., 2005). For instance, Shh signalingregulates the activity of Vax2, which ventralizes the eye by controllingits subcellular localization (Kim and Lemke, 2006). Additionally, anti-Shh antibody treatment results in transdifferentiation of exclusivelythe ventral RPE to NR in chicken (Zhang and Yang, 2001). In summary,it seems that the dorsal RPE development is regulated by Wnt/β-catenin signaling, whereas the ventral RPE development is regulatedby other signaling pathways such as Shh (Fig. 9B).

We have shown that ectopic activation of Wnt/β-catenin signalingresults in the disruption of normal RPE development. At E10.5Wnt/β-catenin signaling was activated in the entire RPE of the gain-of-function mutant and the presumptive RPE appeared to be morpho-logically different compared to that of the wild-type. However, thebasic patterning of RPE and NR was still normal according to theexpression of NR- and RPE-specific markers. Only by E11.5 Mitf andOtx2 became downregulated in the proximal RPE. However, incontrast to the loss-of-function mutant, the expression of Chx10 andRx remained restricted to NR. The expression pattern of these keytranscription factors in the gain-of-function mutant indicates that theproximal RPE development was more seriously disrupted. Apart fromMitf and Otx2, transcription factor Pax6 was downregulated in theproximal RPE, while its expression in the distal RPE remainedunchanged in the gain-of-function mutant. The downregulation ofthe Pax6 gene is particularly interesting because Pax6 together withPax2 is essential for the RPE development and was previously shownto bind toMitf-A promoter and activate its expression in vitro (Baumeret al., 2003). Furthermore, ectopic expression of Pax6 in the optic stalkresults in formation of Mitf-positive RPE (Baumer et al., 2003). It was

recently shown that two other isoforms,Mitf-H andMitf-D, play a keyrole in the RPE development (Bharti et al., 2008). No data is availableabout the possible role of Pax6 in the activation of Mitf-H and Mitf-D.It is nevertheless intriguing to think that the downregulation of Mitfexpression could be caused, at least in part, by Pax6 downregulation inthe proximal RPE.

Taken together, spatial and temporal regulation of Wnt/β-cateninsignaling is essential for the proper RPE development in mice. Wesuggest that the two distinct roles of Wnt/β-catenin signaling in theRPE development could be to regulate differentiation of the RPE byinitiating or maintaining expression ofMitf-D,Mitf-H and Otx2, and toprotect the RPE from NR-specifying signals.

Acknowledgments

We are grateful to Jitka Lachova for maintenance of mice and kindhelp with the Trp1-Cre/Rosa26R experiment, Jana Ruzickova for helpwith Figure presentation, Jan Lukas, Sarka Takacova, Ondrej Machonand Romana Kucerova for valuable comments on the manuscript. Wewish to thank Dr. R. Kemler for Catnb(ex2–6), Dr. S. Piccolo for BAT-galreporter mice, Dr. H. Arnheiter for Mitf antibody, Dr. F. Vaccarino forOtx2 antibody, and Dr. G Lemke for Vax2 probe. This work wassupported by grant no. 204/08/1618 from the Grant Agency of CzechRepublic, by grants nos. 1M6837805002 and B06077 from theMinistry of Education, Youth and Sports of the Czech Republic. TheInstitute of Molecular Genetics is supported by project AV0Z50520514from the Academy of Sciences of the Czech Republic.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ydbio.2009.07.002.

References

Aksan, I., Goding, C.R., 1998. Targeting the microphthalmia basic helix-loop-helix-leucine zipper transcription factor to a subset of E-box elements in vitro and in vivo.Mol. Cell. Biol. 18, 6930–6938.

Allen, B.L., Tenzen, T., McMahon, A.P., 2007. The Hedgehog-binding proteins Gas1 andCdo cooperate to positively regulate Shh signaling during mouse development.Genes Dev. 21, 1244–1257.

Amae, S., Fuse, N., Yasumoto, K., Sato, S., Yajima, I., Yamamoto, H., Udono, T., Durlu, Y.K.,Tamai, M., Takahashi, K., Shibahara, S., 1998. Identification of a novel isoform ofmicrophthalmia-associated transcription factor that is enriched in retinal pigmentepithelium. Biochem. Biophys. Res. Commun. 247, 710–715.

Ashery-Padan, R., Marquardt, T., Zhou, X., Gruss, P., 2000. Pax6 activity in the lensprimordium is required for lens formation and for correct placement of a singleretina in the eye. Genes Dev. 14, 2701–2711.

Barbieri, A.M., Lupo, G., Bulfone, A., Andreazzoli, M., Mariani, M., Fougerousse, F.,Consalez, G.G., Borsani, G., Beckmann, J.S., Barsacchi, G., Ballabio, A., Banfi, S., 1999. Ahomeobox gene, vax2, controls the patterning of the eye dorsoventral axis. Proc.Natl. Acad. Sci. U. S. A. 96, 10729–10734.

Baumer, N., Marquardt, T., Stoykova, A., Spieler, D., Treichel, D., Ashery-Padan, R., Gruss,P., 2003. Retinal pigmented epithelium determination requires the redundantactivities of Pax2 and Pax6. Development 130, 2903–2915.

Beermann, F., Schmid, E., Schutz, G., 1992. Expression of the mouse tyrosinase geneduring embryonic development: recapitulation of the temporal regulation intransgenic mice. Proc. Natl. Acad. Sci. U. S. A. 89, 2809–2813.

Beland, M., Lohnes, D., 2005. Chicken ovalbumin upstream promoter-transcriptionfactor members repress retinoic acid-induced Cdx1 expression. J. Biol. Chem. 280,13858–13862.

Bharti, K., Nguyen, M.T., Skuntz, S., Bertuzzi, S., Arnheiter, H., 2006. The other pigmentcell: specification and development of the pigmented epithelium of the vertebrateeye. Pigment Cell Res. 19, 380–394.

Bharti, K., Liu, W., Csermely, T., Bertuzzi, S., Arnheiter, H., 2008. Alternative promoter usein eye development: the complex role and regulation of the transcription factorMITF. Development 135, 1169–1178.

Brault, V., Moore, R., Kutsch, S., Ishibashi, M., Rowitch, D.H., McMahon, A.P., Sommer, L.,Boussadia, O., Kemler, R., 2001. Inactivation of the beta-catenin gene by Wnt1-Cre-mediated deletion results in dramatic brainmalformation and failure of craniofacialdevelopment. Development 128, 1253–1264.

Burmeister, M., Novak, J., Liang, M.Y., Basu, S., Ploder, L., Hawes, N.L., Vidgen, D., Hoover,F., Goldman, D., Kalnins, V.I., Roderick, T.H., Taylor, B.A., Hankin, M.H., McInnes, R.R.,1996. Ocular retardation mouse caused by Chx10 homeobox null allele: impairedretinal progenitor proliferation and bipolar cell differentiation. Nat. Genet. 12,376–384.

44 N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

Burns, C.J., Zhang, J., Brown, E.C., Van Bibber, A.M., Van Es, J., Clevers, H., Ishikawa, T.O.,Taketo, M.M., Vetter, M.L., Fuhrmann, S., 2008. Investigation of Frizzled-5 duringembryonic neural development in mouse. Dev. Dyn. 237, 1614–1626.

Butler, A.J., Parker, M.G., 1995. COUP-TF II homodimers are formed in preference toheterodimers with RXR alpha or TR beta in intact cells. Nucleic Acids Res. 23,4143–4150.

Cai, Y., Brophy, P.D., Levitan, I., Stifani, S., Dressler, G.R., 2003. Groucho suppresses Pax2transactivation by inhibition of JNK-mediated phosphorylation. EMBO J. 22,5522–5529.

Chen, C.M., Cepko, C.L., 2000. Expression of Chx10 and Chx10-1 in the developingchicken retina. Mech. Dev. 90, 293–297.

Cho, S.H., Cepko, C.L., 2006. Wnt2b/beta-catenin-mediated canonical Wnt signalingdetermines the peripheral fates of the chick eye. Development 133, 3167–3177.

Chow, R.L., Lang, R.A., 2001. Early eye development in vertebrates. Annu. Rev. Cell Dev.Biol. 17, 255–296.

Cooney, A.J., Tsai, S.Y., O'Malley, B.W., Tsai, M.J., 1992. Chicken ovalbumin upstreampromoter transcription factor (COUP-TF) dimers bind to different GGTCA responseelements, allowing COUP-TF to repress hormonal induction of the vitamin D3,thyroid hormone, and retinoic acid receptors. Mol. Cell. Biol. 12, 4153–4163.

DasGupta, R., Fuchs, E., 1999. Multiple roles for activated LEF/TCF transcriptioncomplexes during hair follicle development and differentiation. Development 126,4557–4568.

Dorsky, R.I., Raible, D.W., Moon, R.T., 2000. Direct regulation of nacre, a zebrafish MITFhomolog required for pigment cell formation, by the Wnt pathway. Genes Dev. 14,158–162.

Dorsky, R.I., Sheldahl, L.C., Moon, R.T., 2002. A transgenic Lef1/beta-catenin-dependentreporter is expressed in spatially restricted domains throughout zebrafishdevelopment. Dev. Biol. 241, 229–237.

Dorval, K.M., Bobechko, B.P., Fujieda, H., Chen, S., Zack, D.J., Bremner, R., 2006. CHX10targets a subset of photoreceptor genes. J. Biol. Chem. 281, 744–751.

Duncan, M.K., Haynes Jr., J.I., Cvekl, A., Piatigorsky, J., 1998. Dual roles for Pax-6: atranscriptional repressor of lens fiber cell-specific beta-crystallin genes. Mol. Cell.Biol. 18, 5579–5586.

Filali, M., Cheng, N., Abbott, D., Leontiev, V., Engelhardt, J.F., 2002. Wnt-3A/beta-cateninsignaling induces transcription from the LEF-1 promoter. J. Biol. Chem. 277,33398–33410.

Fu, X., Sun, H., Klein, W.H., Mu, X., 2006. Beta-catenin is essential for lamination but notneurogenesis in mouse retinal development. Dev. Biol. 299, 424–437.

Fuhrmann, S., Levine, E.M., Reh, T.A., 2000. Extraocular mesenchyme patterns the opticvesicle during early eye development in the embryonic chick. Development 127,4599–4609.

Fujimura, N., Vacik, T., Machon, O., Vlcek, C., Scalabrin, S., Speth, M., Diep, D., Krauss, S.,Kozmik, Z., 2007. Wnt-mediated down-regulation of Sp1 target genes by atranscriptional repressor Sp5. J. Biol. Chem. 282, 1225–1237.

Goding, C.R., 2000. Mitf from neural crest to melanoma: signal transduction andtranscription in the melanocyte lineage. Genes Dev. 14, 1712–1728.

Grainger, R.M., Mannion, J.E., Cook Jr., T.L., Zygar, C.A., 1997. Defining intermediate stagesin cell determination: acquisition of a lens-forming bias in head ectoderm duringlens determination. Dev. Genet. 20, 246–257.

Grigoryan, T., Wend, P., Klaus, A., Birchmeier, W., 2008. Deciphering the function ofcanonical Wnt signals in development and disease: conditional loss- and gain-of-function mutations of beta-catenin in mice. Genes Dev. 22, 2308–2341.

Grindley, J.C., Davidson, D.R., Hill, R.E., 1995. The role of Pax-6 in eye and nasal dev-elopment. Development 121, 1433–1442.

Hallsson, J.H., Haflidadottir, B.S., Schepsky, A., Arnheiter, H., Steingrimsson, E., 2007.Evolutionary sequence comparison of the Mitf gene reveals novel conserveddomains. Pigment Cell Res. 20, 185–200.

Harada, N., Tamai, Y., Ishikawa, T., Sauer, B., Takaku, K., Oshima, M., Taketo, M.M., 1999.Intestinal polyposis in mice with a dominant stable mutation of the beta-cateningene. Embo J. 18, 5931–5942.

Hemesath, T.J., Steingrimsson, E., McGill, G., Hansen, M.J., Vaught, J., Hodgkinson, C.A.,Arnheiter, H., Copeland, N.G., Jenkins, N.A., Fisher, D.E., 1994. Microphthalmia, acritical factor in melanocyte development, defines a discrete transcription factorfamily. Genes Dev. 8, 2770–2780.

Hershey, C.L., Fisher, D.E., 2005. Genomic analysis of the microphthalmia locus andidentification of the MITF-J/Mitf-J isoform. Gene 347, 73–82.

Hodgkinson, C.A., Moore, K.J., Nakayama, A., Steingrimsson, E., Copeland, N.G., Jenkins,N.A., Arnheiter, H., 1993. Mutations at the mouse microphthalmia locus areassociated with defects in a gene encoding a novel basic-helix-loop-helix-zipperprotein. Cell 74, 395–404.

Horsford, D.J., Nguyen, M.T., Sellar, G.C., Kothary, R., Arnheiter, H., McInnes, R.R., 2005.Chx10 repression of Mitf is required for the maintenance of mammalianneuroretinal identity. Development 132, 177–187.

Hovanes, K., Li, T.W., Munguia, J.E., Truong, T., Milovanovic, T., Lawrence Marsh, J.,Holcombe, R.F., Waterman, M.L., 2001. Beta-catenin-sensitive isoforms of lym-phoid enhancer factor-1 are selectively expressed in colon cancer. Nat. Genet. 28,53–57.

Huelsken, J., Vogel, R., Brinkmann, V., Erdmann, B., Birchmeier, C., Birchmeier, W., 2000.Requirement for beta-catenin in anterior-posterior axis formation in mice. J. CellBiol. 148, 567–578.

Jasoni, C., Hendrickson, A., Roelink, H., 1999. Analysis of chicken Wnt-13 expressiondemonstrates coincidence with cell division in the developing eye and is consistentwith a role in induction. Dev. Dyn. 215, 215–224.

Jho, E.H., Zhang, T., Domon, C., Joo, C.K., Freund, J.N., Costantini, F., 2002. Wnt/beta-catenin/Tcf signaling induces the transcription of Axin2, a negative regulator of thesignaling pathway. Mol. Cell. Biol. 22, 1172–1183.

Jin, E.J., Burrus, L.W., Erickson, C.A., 2002. The expression patterns of Wnts and theirantagonists during avian eye development. Mech. Dev. 116, 173–176.

Junghans, D., Hack, I., Frotscher, M., Taylor, V., Kemler, R., 2005. Beta-catenin-mediatedcell-adhesion is vital for embryonic forebrain development. Dev. Dyn. 233, 528–539.

Kim, J.W., Lemke, G., 2006. Hedgehog-regulated localization of Vax2 controls eyedevelopment. Genes Dev. 20, 2833–2847.

Klaus, A., Birchmeier, W., 2008. Wnt signalling and its impact on development andcancer. Nat. Rev. Cancer 8, 387–398.

Kreslova, J., Machon, O., Ruzickova, J., Lachova, J., Wawrousek, E.F., Kemler, R., Krauss, S.,Piatigorsky, J., Kozmik, Z., 2007. Abnormal lens morphogenesis and ectopic lensformation in the absence of beta-catenin function. Genesis 45, 157–168.

Kubo, F., Takeichi, M., Nakagawa, S., 2003. Wnt2b controls retinal cell differentiation atthe ciliary marginal zone. Development 130, 587–598.

Kurokawa, D., Kiyonari, H., Nakayama, R., Kimura-Yoshida, C., Matsuo, I., Aizawa, S.,2004. Regulation of Otx2 expression and its functions in mouse forebrain andmidbrain. Development 131, 3319–3331.

Lang, D., Lu, M.M., Huang, L., Engleka, K.A., Zhang, M., Chu, E.Y., Lipner, S., Skoultchi, A.,Millar, S.E., Epstein, J.A., 2005. Pax3 functions at a nodal point in melanocyte stemcell differentiation. Nature 433, 884–887.

Lee, C.S., May, N.R., Fan, C.M., 2001. Transdifferentiation of the ventral retinal pigmentedepithelium to neural retina in the growth arrest specific gene 1 mutant. Dev. Biol.236, 17–29.

Liu, H., Mohamed, O., Dufort, D., Wallace, V.A., 2003. Characterization of Wnt signalingcomponents and activation of theWnt canonical pathway in themurine retina. Dev.Dyn. 227, 323–334.

Liu, H., Thurig, S., Mohamed, O., Dufort, D., Wallace, V.A., 2006. Mapping canonical Wntsignaling in the developing and adult retina. Invest. Ophthalmol. Vis. Sci. 47,5088–5097.

Liu, H., Xu, S., Wang, Y., Mazerolle, C., Thurig, S., Coles, B.L., Ren, J.C., Taketo, M.M., vander Kooy, D., Wallace, V.A., 2007. Ciliary margin transdifferentiation from neuralretina is controlled by canonical Wnt signaling. Dev. Biol. 308, 54–67.

Logan, C.Y., Nusse, R., 2004. The Wnt signaling pathway in development and disease.Annu. Rev. Cell Dev. Biol. 20, 781–810.

Maretto, S., Cordenonsi, M., Dupont, S., Braghetta, P., Broccoli, V., Hassan, A.B., Volpin, D.,Bressan, G.M., Piccolo, S., 2003. MappingWnt/beta-catenin signaling during mousedevelopment and in colorectal tumors. Proc. Natl. Acad. Sci. U. S. A. 100, 3299–3304.

Marquardt, T., Ashery-Padan, R., Andrejewski, N., Scardigli, R., Guillemot, F., Gruss, P.,2001. Pax6 is required for the multipotent state of retinal progenitor cells. Cell 105,43–55.

Martinelli, D.C., Fan, C.M., 2007. Gas1 extends the range of Hedgehog action byfacilitating its signaling. Genes Dev. 21, 1231–1243.

Martinez-Morales, J.R., Signore, M., Acampora, D., Simeone, A., Bovolenta, P., 2001. Otxgenes are required for tissue specification in the developing eye. Development 128,2019–2030.

Martinez-Morales, J.R., Dolez, V., Rodrigo, I., Zaccarini, R., Leconte, L., Bovolenta, P., Saule,S., 2003. OTX2 activates the molecular network underlying retina pigment epithe-lium differentiation. J. Biol. Chem. 278, 21721–21731.

Martinez-Morales, J.R., Rodrigo, I., Bovolenta, P., 2004. Eye development: a view fromthe retina pigmented epithelium. Bioessays 26, 766–777.

Mathers, P.H., Grinberg, A., Mahon, K.A., Jamrich, M., 1997. The Rx homeobox gene isessential for vertebrate eye development. Nature 387, 603–607.

McCaffery, P., Wagner, E., O'Neil, J., Petkovich, M., Drager, U.C., 1999. Dorsal and ventralretinal territories defined by retinoic acid synthesis, break-down and nuclearreceptor expression. Mech. Dev. 82, 119–130.

Miller, L.A., Smith, A.N., Taketo, M.M., Lang, R.A., 2006. Optic cup and facial patterningdefects in ocular ectoderm beta-catenin gain-of-function mice. BMC Dev. Biol. 6, 14.

Mohamed, O.A., Clarke, H.J., Dufort, D., 2004. Beta-catenin signaling marks theprospective site of primitive streak formation in the mouse embryo. Dev. Dyn.231, 416–424.

Mori, M., Metzger, D., Garnier, J.M., Chambon, P., Mark, M., 2002. Site-specific somaticmutagenesis in the retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci. 43,1384–1388.

Mui, S.H., Kim, J.W., Lemke, G., Bertuzzi, S., 2005. Vax genes ventralize the embryoniceye. Genes Dev. 19, 1249–1259.

Muller, F., Rohrer, H., Vogel-Hopker, A., 2007. Bone morphogenetic proteins specify theretinal pigment epithelium in the chick embryo. Development 134, 3483–3493.

Naka, H., Nakamura, S., Shimazaki, T., Okano, H., 2008. Requirement for COUP-TFI and IIin the temporal specification of neural stem cells in CNS development. Nat.Neurosci. 11, 1014–1023.

Nguyen, M., Arnheiter, H., 2000. Signaling and transcriptional regulation in earlymammalian eye development: a link between FGF and MITF. Development 127,3581–3591.

Nornes, H.O., Dressler, G.R., Knapik, E.W., Deutsch, U., Gruss, P., 1990. Spatially and tem-porally restricted expression of Pax2 during murine neurogenesis. Development 109,797–809.

Pereira, F.A., Qiu, Y., Zhou, G., Tsai, M.J., Tsai, S.Y., 1999. The orphan nuclear receptorCOUP-TFII is required for angiogenesis and heart development. Genes Dev. 13,1037–1049.

Pereira, F.A., Tsai, M.J., Tsai, S.Y., 2000. COUP-TF orphan nuclear receptors in develop-ment and differentiation. Cell. Mol. Life Sci. 57, 1388–1398.

Perez-Moreno, M., Jamora, C., Fuchs, E., 2003. Sticky business: orchestrating cellularsignals at adherens junctions. Cell 112, 535–548.

Posthaus, H., Williamson, L., Baumann, D., Kemler, R., Caldelari, R., Suter, M.M., Schwarz,H., Muller, E., 2002. beta-Catenin is not required for proliferation and differentiationof epidermal mouse keratinocytes. J. Cell. Sci. 115, 4587–4595.

Rowan, S., Chen, C.M., Young, T.L., Fisher, D.E., Cepko, C.L., 2004. Transdifferentiation of

45N. Fujimura et al. / Developmental Biology 334 (2009) 31–45

the retina into pigmented cells in ocular retardation mice defines a new function ofthe homeodomain gene Chx10. Development 131, 5139–5152.

Schepsky, A., Bruser, K., Gunnarsson, G.J., Goodall, J., Hallsson, J.H., Goding, C.R.,Steingrimsson, E., Hecht, A., 2006. The microphthalmia-associated transcriptionfactor Mitf interacts with beta-catenin to determine target gene expression. Mol.Cell. Biol. 26, 8914–8927.

Schulte, D., Furukawa, T., Peters, M.A., Kozak, C.A., Cepko, C.L., 1999. Misexpression ofthe Emx-related homeobox genes cVax and mVax2 ventralizes the retina andperturbs the retinotectal map. Neuron 24, 541–553.

Schwarz, M., Cecconi, F., Bernier, G., Andrejewski, N., Kammandel, B., Wagner, M., Gruss,P., 2000. Spatial specification of mammalian eye territories by reciprocaltranscriptional repression of Pax2 and Pax6. Development 127, 4325–4334.

Simeone, A., Acampora, D., Gulisano, M., Stornaiuolo, A., Boncinelli, E., 1992. Nestedexpression domains of four homeobox genes in developing rostral brain. Nature358, 687–690.

Smith, A.N., Miller, L.A., Song, N., Taketo, M.M., Lang, R.A., 2005. The duality of beta-catenin function: a requirement in lens morphogenesis and signaling suppressionof lens fate in periocular ectoderm. Dev. Biol. 285, 477–489.

Soriano, P., 1999. Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat.Genet. 21, 70–71.

Steingrimsson, E., Copeland, N.G., Jenkins, N.A., 2004. Melanocytes and the micro-phthalmia transcription factor network. Annu. Rev. Genet. 38, 365–411.

Takeda, K., Yasumoto, K., Takada, R., Takada, S., Watanabe, K., Udono, T., Saito, H.,Takahashi, K., Shibahara, S., 2000. Induction of melanocyte-specific microphthalmia-associated transcription factor by Wnt-3a. J. Biol. Chem. 275, 14013–14016.

Takeda, K., Yasumoto, K., Kawaguchi, N., Udono, T., Watanabe, K., Saito, H., Takahashi, K.,Noda, M., Shibahara, S., 2002. Mitf-D, a newly identified isoform, expressed in theretinal pigment epithelium and monocyte-lineage cells affected by Mitf mutations.Biochim. Biophys. Acta 1574, 15–23.

Torres, M., Gomez-Pardo, E., Gruss, P., 1996. Pax2 contributes to inner ear patterning andoptic nerve trajectory. Development 122, 3381–3391.

Valenta, T., Lukas, J., Doubravska, L., Fafilek, B., Korinek, V., 2006. HIC1 attenuates Wntsignaling by recruitment of TCF-4 and beta-catenin to the nuclear bodies. EMBO J.25, 2326–2337.

Van Raay, T.J., Moore, K.B., Iordanova, I., Steele, M., Jamrich, M., Harris, W.A., Vetter, M.L.,2005. Frizzled 5 signaling governs the neural potential of progenitors in thedeveloping Xenopus retina. Neuron 46, 23–36.

Veien, E.S., Rosenthal, J.S., Kruse-Bend, R.C., Chien, C.B., Dorsky, R.I., 2008. CanonicalWnt signaling is required for the maintenance of dorsal retinal identity.Development 135, 4101–4111.

West-Mays, J.A., Zhang, J., Nottoli, T., Hagopian-Donaldson, S., Libby, D., Strissel, K.J.,Williams, T., 1999. AP-2alpha transcription factor is required for early morpho-genesis of the lens vesicle. Dev. Biol. 206, 46–62.

Xu, L., Overbeek, P.A., Reneker, L.W., 2002. Systematic analysis of E-, N- and P-cadherinexpression in mouse eye development. Exp. Eye Res. 74, 753–760.

Yan, D., Wiesmann, M., Rohan, M., Chan, V., Jefferson, A.B., Guo, L., Sakamoto, D.,Caothien, R.H., Fuller, J.H., Reinhard, C., Garcia, P.D., Randazzo, F.M., Escobedo, J.,Fantl, W.J., Williams, L.T., 2001. Elevated expression of axin2 and hnkd mRNAprovides evidence that Wnt/beta -catenin signaling is activated in human colontumors. Proc. Natl. Acad. Sci. U. S. A. 98, 14973–14978.

Yasumoto, K., Yokoyama, K., Shibata, K., Tomita, Y., Shibahara, S., 1994. Microphthalmia-associated transcription factor as a regulator for melanocyte-specific transcriptionof the human tyrosinase gene. Mol. Cell. Biol. 14, 8058–8070.

Yasumoto, K., Yokoyama, K., Takahashi, K., Tomita, Y., Shibahara, S., 1997. Functionalanalysis of microphthalmia-associated transcription factor in pigment cell-specific transcription of the human tyrosinase family genes. J. Biol. Chem. 272,503–509.

Zakin, L.D., Mazan, S., Maury, M., Martin, N., Guenet, J.L., Brulet, P., 1998. Structureand expression of Wnt13, a novel mouse Wnt2 related gene. Mech. Dev. 73,107–116.

Zhang, X.M., Yang, X.J., 2001. Temporal and spatial effects of Sonic hedgehog signaling inchick eye morphogenesis. Dev. Biol. 233, 271–290.

Zhou, C.J., Molotkov, A., Song, L., Li, Y., Pleasure, D.E., Pleasure, S.J., Wang, Y.Z., 2008.Ocular coloboma and dorsoventral neuroretinal patterning defects in Lrp6 mutanteyes. Dev. Dyn. 237, 3681–3689.

Zhou, J., Qu, J., Yi, X.P., Graber, K., Huber, L., Wang, X., Gerdes, A.M., Li, F., 2007.Upregulation of gamma-catenin compensates for the loss of beta-catenin in adultcardiomyocytes. Am. J. Physiol. Heart Circ. Physiol. 292, H270–H276.